Subgroup Analysis — SCE Medical Oncology MCQ

Correct answer: C — There is no evidence of effect modification; the imprecise subgroup estimate should not be interpreted alone

Explanation lettering: E = shown as B · B = shown as E

C is correct. The non-significant interaction provides no evidence that the relative treatment effect differs between this subgroup and its complement. The subgroup confidence interval is wide and crosses 1, showing limited precision rather than proving absence of effect. A commits the difference-in-significance error. B ignores uncertainty. D reverses interpretation of the interaction p value. E treats an expected loss of power in a smaller subset as invalidation of the overall trial. Credibility also depends on prespecification, multiplicity, biological plausibility, complement-subgroup results and consistency across studies, not one isolated p value.

Reference: EMA guideline on subgroups in confirmatory trials: https://www.ema.europa.eu/en/investigation-subgroups-confirmatory-clinical-trials-scientific-guideline